Research Of Wind Generator Condition Monitoring Based On Information Fusion

In recent years, with the rapid development of wind power industry, the installed capacity and the units quantity of the wind power plant keep improving. Meanwhile, the fault of wind generator is also increasing. Improve the wind power generation technology, while ignoring monitoring of running status of wind turbines, which lead to major failure of the generator. This situation not only affected the overall wind farm operation and economic benefit, but easy to cause safety accidents. The traditional monitoring methods can not fully and effectively response machine running status, in this paper, on the basis of the further study of information fusion technology, applied it to the wind generator condition monitoring, in order to accurately identify the normal or abnormal state of the wind turbines.This paper studies the internal structure and working mode of wind generator, reference for large rotating machinery, analyze the common fault of its main transmission system in the method of vibration. As the complex structure and bad working condition of the megawatt wind generator, the traditional monitoring method can’t get good effect, so put in the information fusion technology, using multi-sensor detection, fusing the data feature to monitor the wind generator. Set up the information fusion model, process the vibration signal collected by the single sensor, do the time domain, frequency domain and the related analysis. Research information fusion algorithm, and put forward an improved bayesian method, input the feature of signal into the model to do the feature level fusion, get the evaluation result of single sensor. Then input every result of the sensors into D-S evidence theory fusion model, get the final identify result of the wind generator condition. Develop the monitoring system on the platform of virtual instrument Labview, using the matlab script node to conduct the combined programming with Labview and MATLAB, give full play to the advantages of two software in their respective fields, complete model building, data fusion, etc.According to the Wind generator vibration condition monitoring guideline (NB/T31004-2011) published by National Energy Administration and the German standard of wind turbines (VDI3834-1-2009), combined with the practical situation of the wind generator, determine the position and installation of the sensor. Tested the designed system in lab and wind power plant, and turned out that the wind turbine condition monitoring system based on information fusion technology designed in this paper is effective in the analysis of signal processing and accurately identified the running state of the wind generator.